troubleshooting
Fail
Audited by Snyk on Nov 21, 2026
Risk Level: CRITICAL
Full Analysis
CRITICAL E004: Prompt injection detected in skill instructions.
- Potential prompt injection detected (medium risk: 0.30). The skill is a standard Kubernetes troubleshooting guide with no obfuscation, pretexts, or covert behavior, but it includes unrequested cluster-mutating commands (a deployment resource change and a cluster-wide certificate renewal) given without warnings or confirmation steps, which is a mild risk under the recall-first override.
MEDIUM W012: Unverifiable external dependency detected (runtime URL that controls agent).
- Potentially malicious external URL detected (medium risk: 0.30). The skill's only external inputs are container image references run via
kubectl run/debug, and thenicolaka/netshootcommunity image is a repository locator executed in-cluster from an individual-maintained project (unofficial but reputable), so it is a medium-tier dependency, while busybox (Docker Official Image) and the gcr.io Kubernetes e2e dnsutils image are first-party and strong.
MEDIUM W013: Attempt to modify system services in skill instructions.
- Attempt to modify system services in skill instructions detected (low risk: 0.10). The skill exposes state-changing cluster operations (
kubectl set resourceson a Deployment andkubeadm certs renew all), but no explicit irreversible operational loss is established, so the tier is capped at low.
Issues (3)
E004
CRITICALPrompt injection detected in skill instructions.
W012
MEDIUMUnverifiable external dependency detected (runtime URL that controls agent).
W013
MEDIUMAttempt to modify system services in skill instructions.
Audit Metadata